DeepSeek V4.1 Flash is free on OpenCode Zen for a limited time with 200K context
OpenCode Zen added a free version of DeepSeek V4.1 Flash. Input, output, and cached reads cost nothing for a limited time. The free version has about 200K tokens of context and up to about 32K output tokens, supports image input and tool calls, and its provider does not store or train on your requests. No end date has been announced.
The briefing
OpenCode Zen now has a free version of DeepSeek V4.1 Flash (model id deepseek-v4.1-flash-free). Input, output, and cached reads are all billed at $0 for a limited time, and no end date has been announced.
What you can use
The free version is the same DeepSeek V4.1 Flash model, with reasoning, tool calls, structured output, and image input. The difference is length: the free version has a context window of about 200K tokens and up to about 32K output tokens, while the paid deepseek-v4.1-flash has a 1M context window and costs $0.30 per million input tokens and $1.20 per million output tokens on Zen.
The provider serving the free version follows a zero-retention policy, so your requests are not stored or used for training. 200K context covers everyday coding and small agent tasks; switch back to the paid version when you need to load a large repo in one go.
The free version is only on Zen for now. DeepSeek V4.1 Flash inside an OpenCode Go plan still counts against your Go usage limits.
How to use it
- Open OpenCode, sign in to Zen, and switch to DeepSeek V4.1 Flash Free with
/models, or set the model toopencode/deepseek-v4.1-flash-freein your config. - For API calls, use the Zen endpoint
https://opencode.ai/inference/v1/chat/completionswith the model namedeepseek-v4.1-flash-free.
More details: OpenCode Zen docs.